You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Downside: usually it's up to the feed provider to solve invalid feed issues which might happen or not, so validation issues might clutter the log, potentially hiding issues with the lamassu instance itself.
The text was updated successfully, but these errors were encountered:
Or set the level to warning and document how logback could be configured to filter out specific log messages in case they would pollute the log. That way, a very fine grained config would be possible, e.g. just filtering out warnings for well known systems.
I'm not really sure how this would have helped in the example case. We wouldn't have known about it before it happened, and when it happened the problem ended up in an exception. Imho looking at validation metrics or alerting is more useful than scanning logs for these issues.
If you need these though, I think there's already a way to override the log level via properties at class level:
org.entur.lamassu.leader.FeedUpdater.level=WARN
As for logback filtering, I'm not familiar with it, but if you provide some examples and sane defaults, I might consider it.
Currently, feed validation errors are logged at log level
INFO
. As invalid feeds might cause unexpected results in downstream processing (as e.g. missing vehicle_types_available currenty does), I suggest to rather log them at levelWARNING
.Downside: usually it's up to the feed provider to solve invalid feed issues which might happen or not, so validation issues might clutter the log, potentially hiding issues with the lamassu instance itself.
The text was updated successfully, but these errors were encountered: